PNS219 National and Regional Representativeness of the National Health and Wellness Survey Compared to the Behavioral Risk Factor Surveillance System

Abstract

National health surveys provide insight into population health behaviors and outcomes. To ensure generalizability of findings, surveys are often weighted against reputable sources such as the US Census Bureau (USCB). The purpose of this study is to determine national and regional representativeness of the National Health and Wellness Survey (NHWS) compared to the Behavioral Risk Factor Surveillance System (BRFSS). Data were obtained from 2018 NHWS (n=74,977) and 2018 BRFSS (n=429,836). NHWS is a self-administered, web-based survey of a representative sample of US adults aged 18+, weighted against age, sex, ethnicity, and education (USCB Current Population Survey 2017). BRFSS is a telephone-based survey of noninstitutionalized US adults, weighted against age, sex, race/ethnicity (Claritas, LLC), education, marital status, home-owner/renter-ship (USCB American Community Survey), and telephone wireless percentages (National Center for Health Statistics). National and Regional (Northeast, Midwest, South, West) weighted prevalence estimates were calculated for sex, age, race/ethnicity, education, employment, income, body mass index (BMI), smoking, and exercise status. Significance testing was performed on weighted estimates for continuous (student’s t-test) and categorical (chi-squared test) variables. Most national and regional demographic variables and prevalence estimates differed by <2% between surveys. The following variables were not significantly different: sex (female: 51.8% (NHWS), 51.3% (BRFSS); p=0.0967), Midwestern smoking status (smoker: 17.7% (NHWS), 17.4% (BRFSS); p=0.4981), Southern BMI (28.4±0.053 (NHWS), 28.5±0.041 (BRFSS); p=0.3272), Southern smoking status (smoker: 16.1% (NHWS), 15.9% (BRFSS); p=0.4638), Western sex (female: 49.6% (NHWS), 50.6% (BRFSS); p=0.0944), and Western BMI (27.5±0.067 (NHWS), 27.5±0.043 (BRFSS); p=0.6241). Despite applying weighting estimates from different sources and having different modes of administration (online versus telephone), prevalence estimates were comparable between NHWS and BRFSS. Thus, NHWS is nationally and regionally representative for population demographics and health behaviors. Comparison of prevalence estimates for self-reported physician-diagnoses and healthcare resource utilization will be explored in follow-up analyses.